Genomes evolve and diversify through di?erent mechanisms, including small point mutations, but also larger, structural variations (SV). SVs can be mediated by simple repeats and microhomology based recombination (termed 'progressive SVs' in this proposal). However, progressive SV mechanisms cannot explain all forms of large genomic variation; sometimes, more 'complex mechanisms' are needed; examples include Breakage Fusion Bridge, and Chromothripsis. Moreover, there is little understanding of the genetic mechanisms of genome instability that lead to complex SV formation. It is suspected that random viral genome insertions into the genome can on occasion disrupt key genes, causing genome instability and hyper-variability. To address these problems, the proposal will design and implement computational methods to (a) reconstruct and validate episomal structures of viral genome insertions; (b) determine if genomic sequence sampled from tumor genomes has a signature of complex variation; and, (b) phase and sub-type regions with complex SV including KIR and HLA; As clinical/translational applications of genomics come to the forefront, the impact of complex SVs on the phenotype of an individual become increasingly important. Understanding the computational signatures of BFB and Chromothripsis will help sub-type and characterize cancers. The knowledge of KIR/HLA sub-type will be correlated with immune related phenotypes, and the reconstruction of viral episomes will help clarify the etiology of virus mediated cancers. Thus, the proposed set of computational tools will directly impact the translational/medical aspect of genomics.